If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Increasing resistance of the Liverpool Epidemic Strain (LES) of Pseudomonas aeruginosa (Psa) to antibiotics in cystic fibrosis (CF)—A cause for concern?
Transmissible Pseudomonas aeruginosa (Psa) strains such as the Liverpool Epidemic Strain (LES) are now widespread throughout UK CF clinics: their susceptibility to antibiotics is therefore important. To study this, we compared antibiogram patterns of Psa strains in our CF clinic over 5 years, looking at differences in resistance patterns between strains and changes to these over time.
Methods
The antibiograms of sputum samples between 2004 and 2008 from patients attending our centre were included. We compared Psa isolate antibiotic resistance (to six anti-pseudomonal antibiotics) patterns for patients infected with LES with those infected with other Psa strains, both in the total population in 2004 (125 patients) and 2008 (166 patients) and also longitudinally from annual review samples 2004 to 2008 in matched and unmatched patient groups.
Results
LES exhibited significantly more resistant isolates in 2004 (p<0.0001). There was an increase in antibiotic resistance in both LES and other Psa strains over time (p<0.001). Cox proportional hazards analysis of both unmatched (n=125) and matched (n=56) patients in 2004 revealed that LES infected patients were more likely to develop antibiotic resistant isolates over time (hazard ratio 8.1, p<0.001). Fewer LES isolates were classed as fully sensitive in both matched and unmatched groups at the end of study period (p<0.001).
Conclusion
This study shows a worrying trend in antibiotic resistance in the Psa isolates amongst patients chronically infected with LES. This highlights the need to prevent cross infection through segregation and also the need to develop new strategies to treat these organisms.
]. Treatment of these infections involves combinations of intravenous anti-Psa antibiotics, often given over prolonged periods, which may encourage antibiotic resistance [
]. Furthermore, the recent emergence of Psa clones that are highly transmissible between CF individuals and have been shown to confer a worse prognosis [
Despite this, there have been no systematic studies looking at antibiotic resistance over time in Psa in the CF community. We therefore undertook a 5 year prospective study of the antibiotic resistance of Psa in patients attending a large adult CF unit in the north-west of England, where there is a cohort of individuals chronically infected with the Liverpool Epidemic Strain (LES) of Psa, the most prevalent and important such transmissible strain in the UK [
]. We looked for any differences in resistance patterns between strains and whether these patterns have changed over time.
2. Patients and methods
2.1 Patient cohorts
The Liverpool Adult CF Unit provides specialist services for North Wales, Cheshire, Merseyside, and the Isle of Man: a catchment of approximately 3 million people. All adult CF patients have regular (at least bimonthly) sputum analysis for Psa and to prevent cross-infection we have regularly genotyped their strains for many years.
Patients who were chronically infected with Psa (defined by at least 3 positive sputum samples within a 6 month period) [
] formed the study population. For the purposes of this study, patients additionally chronically infected with Burkholderia cepacia complex and the only 2 with another known transmissible Psa strain (Midlands1) were excluded.
We compared Psa isolate antibiotic resistance patterns for patients infected with LES with those infected with sporadic Psa strains, from similar numbers of sputum samples taken in 2004 (125 patients, 91 LES) and 2008 (166 patients, 110 LES) (Table 2) and also longitudinally from 2004 to 2008 in both matched and unmatched patient groups (see below). For the latter study, to standardise antibiogram selection the results from the sputum sample closest to the anniversary of the patient's birthday were used for each year.
] for LES group membership in order to allow matching with the other group. This was determined without regard to outcome, using multivariable logistic regression analysis with a full non-parsimonious model that included all available patient characteristics, balancing all items recorded that may relate to either systematic bias or simple chance. This model yielded a C statistic of 0.78, indicating an acceptable ability to differentiate between patients with or without LES. We then used a macro (www2.sas.com/proceedings/sugi29/165-29.pdf) to perform propensity-matching. Using this model, in 2004 28 LES infected patients were matched with 28 infected with other Psa strains on clinical parameters (age, sex, FEV1, BMI, time since diagnosis of CF, hospital admissions, outpatient visits and baseline antibiotic resistance) (Table 1). 2004 patient demographics are shown in Table 1.
Table 1Characteristics of 125 adult CF patients chronically infected with P. aeruginosa attending the Liverpool Adult CF Centre in 2004 (unmatched and matched groups).
Unmatched
Matched (n=56)
Other Psa infection (n=34)
LES Psa infection (n=91)
p value
Other Psa infection (n=28)
LES Psa infection (n=28)
p value
Female (%)
43
53
0.44
46
43
0.79
Years since diagnosis
22 (16–27)
25 (23–29)
0.003
23 (18–29)
23 (21–26)
0.62
Age (years)
26 (21–34)
29 (24–36)
<0.01
25 (23–34)
27 (25–32)
0.18
BMI
22 (20–24)
21 (19–23)
0.08
21 (20–24)
22 (19–25)
0.80
FEV1 (%)
64 (46–83)
64 (45–85)
0.88
63 (45~88)
77 (49–94)
0.45
Hospital admissions/year
4 (1–7)
8 (4–14)
<0.001
4 (1–7)
6 (3–10)
0.45
Outpatient visits
33 (21–38)
36 (22–50)
0.29
33 (21–36)
31 (17–49)
0.89
Mean home iv days/year
7.3
7.9
0.10
6.095
6.592
0.77
Mean antibiotic resistance (%)
11.4
46.2
<0.001
14.3
10.7
>0.99
Categorical variables shown as percentage, comparisons made with χ2 test.
Continuous variables shown as median (25th–75th percentiles), comparisons made with Wilcoxon rank sum test.
Cox proportional hazards analysis was then used to determine changes in antibiotic resistance over time in both the matched and unmatched groups.
Statistical significance was calculated using Fisher's exact test and Chi squared tests with Yates correction: p<0.05 was considered to be significant. Stata software version 2.0 was used to analyse the data.
3. Microbiological methods
3.1 Antibiotic susceptibility testing
Sputum samples were treated with equal quantities of sputasol (oxoid), agitated at 200 rpm for 15 min at room temperature, plated onto blood and chocolate agar and MacConkey agar, and incubated at 37 °C. The plates were examined at 48 h for growth of Psa colonies, and those with distinct morphotypes were further cultured onto purity plates (chocolate agar). Each morphologically different subtype (up to 5 per sample) was tested for antibiotic susceptibility using BSAC guidelines [
] by the disc diffusion method using the following: tazobactam/piperacillin 85 μg, meropenem 10 μg, tobramycin10 μg, ciprofloxacin 5 μg, ceftazidime 30 μg and colistin sulphate 25 μg. The sizes of the zones of inhibition (mm) were recorded and susceptibilities were defined as per standard protocols. In total 9200 susceptibility patterns were analysed.
Sputum samples with multiple morphotypes were judged to be sensitive only if all morphotypes were sensitive to at least two of the three classes of antibiotics tested (fluoroquinolones, aminoglycosides, and beta-lactams/carbapenems). Strains were defined as resistant if they were resistant to two of the three classes of antibiotics [
] and pan-resistant if resistant to all three classes of antibiotics. Isolates which exhibited intermediate resistance were considered resistant.
3.2 Psa genotyping
Oligonucleotide primers (Sigma-Genosys) used in PCR assays were (PASS, PS21R, PS21F, LESF9F, LESF9R). DNA for PCR amplification was prepared by making a suspension of a few colonies in 200 ml 5% Chelex-100 (Sigma) solution. After vigorous mixing, the suspension was boiled for 5–10 min. Following centrifugation, 150 ml of supernatant containing the DNA was removed and stored at 220 °C. Typically, 1 ml of this DNA was used directly in 25 μl volumes containing 1.25 unit Taq DNA polymerase (Promega), 1× Taq Master (Helena Biosciences), 300 nM each primer, 1× Taq buffer, 2.5 μM MgCl2 and 100 μM nucleotides (dATP, dCTP, dGTP, dTTP). Amplifications were carried out in an Eppendorf MasterCycler thermal cycler for 30 cycles consisting of 95 °C (1 min), chosen annealing temperature (1 min) and 72 °C (2 min), with an additional extension time at 72 °C (10 min) following completion of the 30 cycles.
Psa and LES strains were identified using gel electrophoresis looking for bands representing specific amplicons [
Environmental contamination with an epidemic strain of Pseudomonas aeruginosa in a Liverpool cystic fibrosis centre, and study of its survival on dry surfaces.
]. The PA-SS, primer pair, specific for all Psa, yields an amplicon of 956 bp; the LES-specific primer pairs LESF9 and PS21 yield amplicons of 461 bp and 364 bp respectively. PCR tests leading to the PA-SS amplicon only were deemed LES-negative. Tests leading to the presence of all three amplicons (PA-SS, LESF9 and PS21) were deemed LES-positive [
Use of suppression subtractive hybridization to examine the accessory genome of the Liverpool cystic fibrosis epidemic strain of Pseudomonas aeruginosa.
4.1 Cross-sectional (unmatched) analysis of 2004 and 2008 cohorts
Analysis of the 2004 antibiograms showed that LES was more resistant to ceftazidime, meropenem, piperacillin/tazobactam (all p<0.001) and tobramycin (p=0.0001) than other Psa strains, which in turn were more resistant to ciprofloxacin (p<0.002) (Table 2). There was no resistance to colomycin in either group. Overall, resistance was greater in the LES group (mean 18.1%) than the other Psa group (mean 7.2%) (p<0.0001).
Table 2Mean antibiotic resistance exhibited by LES Psa and other Psa isolates to six commonly used antibiotics in 2004 and 2008.
Analysis of the 2008 antibiograms showed that, compared to 2004, the overall mean antibiotic resistance increased in both groups, but LES isolates became significantly more resistant (p<0.0001) to all antibiotics except colomycin when compared to other Psa isolates (Table 2). Although in 2008 the other Psa isolates showed increased mean resistance to ciprofloxacin (10.6% vs 20.6%, p=0.0002), tobramycin (3.5 vs 8.5, p=0.002) and colomycin (0 vs. 2.8, p=0.008), the LES isolates had increased mean resistance to all six antibiotics used (p<0.0001, except piperacillin/tazobactam p=0.01). The proportions of resistant and pan-resistant strains were also significantly higher in the LES group (see Fig. 1).
Fig. 1Prevalence of strain resistance in patients infected Psa followed from 2004 to 2008.
The 125 patients in the 2004 cohort were followed for up to 5 years. Compared to the other Psa infected group, in terms of baseline demographics the 91 patients chronically infected with LES had a significantly higher age, mean antibiotic resistance, time since diagnosis, and number of hospital admissions, but no difference in sex, FEV1, BMI or number of outpatient visits than those with other strains (see Table 1). Although Cox proportional hazards analysis revealed that the LES infected patients were more likely to develop antibiotic resistant isolates over time (hazard ratio 8.1, p<0.001) (Table 3), as were female CF patients (hazard ratio 1.56, p=0.045) and those with a lower baseline FEV1 (hazard ratio 0.99, p=0.03), BMI and time since the diagnosis of CF had no effect. At the end of the 5 year study period, there were 15 deaths and 3 transfers in patients with LES compared to 4 deaths and 1 transfer in those with other Psa, and fewer LES group patient isolates were classed as sensitive (9/73), compared to 25/31 in the other Psa group (p<0.001) (see Fig. 1).
Table 3Cox proportional hazards for follow up antibiotic resistance in unmatched and matched cohorts.
Cox proportional hazards analysis of the 56 matched patients showed that LES infection (hazard ratio 6.1, p<0.001) and lower baseline FEV1 (hazard ratio 0.98, p=0.03) increased the likelihood of developing antibiotic resistant isolates over time, whilst female sex, BMI or time since diagnosis had no effect (Table 3). Even when adjusted for those clinical variables predicting antibiotic resistance over time (age, sex, BMI, FEV1, number of hospital admissions), LES infected patients developed more antibiotic resistant isolates during the 5 year period (Fig. 2) (isolates classed as sensitive: LES infected group [21.7%] versus other Psa infected group 22 [88.5%], p<0.001).
Fig. 2Antibiotic sensitivity profile of patients chronically infected with LES and other Psa strains attending Liverpool Adult CF Unit between 2004 and 2008.*
For the first time to our knowledge we report a longitudinal analysis of the antibiotic sensitivity patterns of P. aeruginosa strains in chronically infected adult CF patients, including those of the most common transmissible Psa strain in the UK, LES. Our results demonstrate a high prevalence of antibiotic resistance in Psa isolates in patients infected with LES and an increased rate of acquisition of antibiotic resistance over time, compared to other Psa strains.
Morbidity in patients with CF results from chronic suppurative lung disease due to Psa, where repeated exacerbations cause progressive lung damage leading to respiratory failure and death in the majority of cases [
]. Possibly due to selective antibiotic pressures, over time these organisms undergo genotypic changes leading to the emergence of treatment-resistant phenotypes: it was the appearance of such a ceftazidime-resistant phenotype at a local paediatric CF centre that led to the important discovery of the original transmissible clone of Psa, LES, in 1996 [
] none have demonstrated the presence of an excess of multiresistant strains, and there have been no previous studies looking at how the antibiotic sensitivity patterns of these epidemic strains may change over time. We have a large cohort of LES-infected CF patients, mainly inherited from paediatric practice, and this has given us the opportunity to study this strain over a prolonged period.
From our longitudinal study of 125 adult patients over 5 years, we have demonstrated that LES is not only more resistant to antibiotics, but also has an enhanced rate of acquisition of antibiotic resistance over time. We have previously shown that chronic infection with LES causes increased hospitalisation and IV antibiotic use [
]. However, this increased antibiotic resistance may not simply be a reflection of repeated antibiotic exposure leading to the emergence of resistance by natural selection, since the effect was still seen following adjustment for confounding factors, and antibiotic pressure only had a weak effect. In keeping with accepted practice, we routinely use at least two antipseudomonal drugs of different classes, and base their selection in those with resistant strains, on the results of previous experience of those antibiotics, in consultation with the patient. For those with sensitive strains, we rotate antibiotics to lessen the chances of resistance occurring.
The ability of Psa to mutate rapidly (hypermutate) in the harsh environment of the CF lung that may give it a survival advantage [
], has also been suggested as an explanation for the emergence of antibiotic resistance. However, although hypermutable Psa strains can occur in up to 36% of CF patients [
], suggesting that an alternative explanation for its resistance to antibiotics is required.
Many isolates of LES, including those associated with transmission to non-CF patients, exhibit an unusual quorum sensing (QS) phenotype termed hypervirulence [
]. This phenotype is characterised by a dysfunctional QS system, leading to overproduction of QS-regulated factors too early in the growth phase. These hypervirulent isolates produce high levels of pyocyanin and other QS-regulated exoproducts of relevance to CF infections [
]. The phenotype has also been linked to increased resistance to some antibiotics. Although the mechanism for this is unknown, it may be related to up-regulation of QS-regulated efflux pumps [
]. The hypervirulence phenotype has not been identified in other Psa isolates, whether in CF or other hosts. Hence, it seems feasible that it might play a role in the success of this epidemic clone, the greater morbidity associated with it, and the development of resistance.
LES has also been found to possess an unstable genotype: the subtypes within LES undergo deletions, insertions and rearrangement leading to significant phenotypic changes [
Phenotypic variability of Pseudomonas aeruginosa in sputa from patients with acute infective exacerbation of cystic fibrosis and its impact on the validity of antimicrobial susceptibility testing.
]. In a UK based survey looking at several transmissible strains, LES had greater microheterogeneity than other strains underlining its genetic variability [
There are limitations to our study. Firstly, we applied in vitro antimicrobial sensitivity testing to organisms existing in the special circumstances of the CF lung, where Psa lives in a complex anaerobic bio-film community [
Antibiotic susceptabilities of Pseudomonas aeruginosa isolates derived from patients with cystic fibrosis under aerobic, anaerobic, and biofilm conditions.
], and exhibits extensive phenotypic divergence, including a mucoid form. Removing these organisms and growing them in planktonic media in aerobic conditions may not mirror sufficiently their natural habitat. Furthermore, many different subtypes may exist within the same sputum sample, each with a different antibiogram [
Phenotypic variability of Pseudomonas aeruginosa in sputa from patients with acute infective exacerbation of cystic fibrosis and its impact on the validity of antimicrobial susceptibility testing.
]. Nevertheless, for the longitudinal study we standardised our approach to the assessment of antibiotic resistance by collecting samples at the same fixed time points for each patient, and utilising industry standard methodology [
] towards antimicrobial testing. Furthermore, by using the same technique for all Psa isolates (both LES and other strains) over a prolonged period the results produced are comparable. Secondly, we cannot be sure that the increased antibiotic resistance in the LES group does not simply reflect an increased use of antibiotics prior to the commencement of the study. Since most patients acquired LES in childhood, the length of infection could not be ascertained, and this group had more hospital admissions. It was for this reason that we used matched groups with similar numbers of hospital admissions and outpatient visits for the longitudinal study. Over the study period, in the matched patients LES exhibited a much greater increase in antibiotic resistance than other Psa strains, strongly suggesting that this was not simply due to pre-existing resistance.
Even though individual sputum sensitivities may not be representative of the true sensitivity in an infecting bacterial population, using longitudinal data over 5 years gives us a glimpse of the overall trends of the antibiotic resistance patterns in these patients, confirming the developing problem of antibiotic resistance in transmissible Psa in CF patients, particularly in those with LES infection. Secondly, we only studied LES and not other transmissible strains, so we cannot be sure that our conclusions are valid for all such strains. However, LES is the most prevalent transmissible strain in the UK and has already spread to units elsewhere, so our findings will be of relevance to all CF units.
In conclusion, this study shows a worrying trend in antibiotic resistance in Psa isolates in the CF population, particularly amongst those chronically infected with LES, which not only exhibits more antibiotic resistance but also develops resistance much more rapidly than other Psa strains. This highlights the requirement for new antimicrobial therapies and the paramount need to enforce stringent infection control and segregation policies in CF units to limit the spread of transmissible Psa strains to improve outcomes for CF patients.
Conflict of interest
None.
References
Kerem E.
Corey M.
Gold R.
Levison H.
Pulmonary function and clinical course in patients with cystic fibrosis after pulmonary colonization with Pseudomonas aeruginosa.
Environmental contamination with an epidemic strain of Pseudomonas aeruginosa in a Liverpool cystic fibrosis centre, and study of its survival on dry surfaces.
Use of suppression subtractive hybridization to examine the accessory genome of the Liverpool cystic fibrosis epidemic strain of Pseudomonas aeruginosa.
Antibiotic susceptabilities of Pseudomonas aeruginosa isolates derived from patients with cystic fibrosis under aerobic, anaerobic, and biofilm conditions.
Phenotypic variability of Pseudomonas aeruginosa in sputa from patients with acute infective exacerbation of cystic fibrosis and its impact on the validity of antimicrobial susceptibility testing.
00210-4&id=gr1.jpg){kind=link}
00210-4&id=gr2.jpg){kind=link}